CHAPTER 5. HOME RANGE
This chapter investigates home range and movement behaviour of the desert-dwelling
giraffe in the northern Namib Desert, using both individual identification methods and
GPS satellite tracking.
It also investigates long distance movements and movement
corridors used by giraffe between the three study areas, and the relationship between
movement and temperature.
5.1. Introduction
Mammals generally range over an area that is just large enough to satisfy their ecological
and biological requirements, such as food, water, mating opportunities and refugia. Home
range size scales with bioenergetic needs, which increase proportionally with body size
(McNab, 1963). Home range size to body mass scaling coefficients are substantially
greater for mammals than other terrestrial species (e.g. McNab, 1963; Lindstedt et al.,
1986; Swihart et al., 1988; du Toit, 1990a), reflecting the costs of homeothermy. In
addition, much variation occurs in home range size due to sex, age and migration patterns,
and environmental factors such as habitat structure, competition and resource availability
(e.g. Clutton-Brock et al., 1989; Fischer & Linsenmair, 1999). In many mammals, food
appears to be a particularly important determinant of home range, with both the dispersion
of food patches (e.g. MacDonald, 1983) and quantity of food (e.g. Boutin, 1990)
influencing range limits in different species.
Dispersal and home range sizes of arid-adapted mammals are larger in comparison to the
same or similar species in higher rainfall environments (e.g. Viljoen, 1989a; du Toit, 1990;
Dickman et al., 1995; Le Pendu & Ciofolo, 1999). Local rainfall, availability of food and
free water are dominant factors driving mammal movements in arid environments, with
individuals moving there to obtain patchy and ephemeral resources than in mesic
environments where resources are often richer and more stable.
Chapter 5: Home Range
96
The giraffe, with a large body mass and high bioenergetic requirements, has more
expansive home ranges than smaller ungulates in the same environment, such as kudu or
steenbok (e.g. Cloete & Kok, 1986; du Toit, 1990a). However, large differences in giraffe
home range sizes have been reported across the range of the species, as well as in different
habitats and under varying management regimes (Langman, 1973; Berry, 1978; Dagg &
Foster, 1982; Pratt & Anderson, 1982; Le Pendu & Ciofolo, 1999; van der Jeugd & Prins,
2000). Abiotic (e.g. climate, topography), biotic (e.g. forage availability and quality,
herbivore and predator densities) and human influences (e.g. poaching, settlements, fences)
are contributing factors that affect giraffe range and distribution (e.g. Dagg & Foster, 1982;
Ciofolo, 1995).
Niger’s desert-dwelling giraffe range individually over a substantially larger area (127 to
1,559 km2; Le Pendu & Ciofolo, 1999) than any other giraffe in Africa; for example, 0.1 to
27 km2 in Tanzania (van der Jeugd & Prins, 2000); 145 km2 in Zambia (Berry, 1978); and
5 to 654 km2 in South Africa (Langman, 1973). Increased vegetation density in savanna
environments increases the available forage per area unit. This correlates, in turn, with
smaller home ranges of giraffe that occur in such habitats (e.g. Langman, 1973; Pratt &
Anderson, 1982; van der Jeugd & Prins, 2000). This relationship supports du Toit’s
(1990a) theory that food production sets the minimal home range limit in mammals.
The use of different sampling techniques and poor seasonal accessibility for researchers
has often resulted in underestimation of giraffe homes (Langman, 1973; Berry, 1978; Dagg
& Foster, 1982; Scheepers, 1992; van der Jeugd & Prins, 2000). Even today, most home
range analyses depend on field observations of individuals. Advanced methods of animal
movement and core home range analysis (Burgman & Fox, 2003; Douglas-Hamilton et al.,
In press) are providing more accurate estimates of animal ranges and movements.
However, the lack of long-term studies remains the most limiting factor in understanding
the home range and movements of giraffe.
Equipping animals with radio transmitter collars has aided ecological research by allowing
remote collection of data on animal movements and home ranges. Radio telemetry studies
have been undertaken on numerous large mammal species, including caribou (e.g. Simpson
et al., 1997), elk, mule and black-tailed deer (e.g. Gillingham & Bunnell, 1985; Harestad,
Chapter 5: Home Range
97
1985; Edge & Marcum, 1989), lions and leopards (e.g. Stander & Hansen, 2003) and
grizzly bears (e.g. Hamilton & Archibald, 1985). More recently, advances in satellite
technology have enabled the compaction of GPS transmitters onto collars. The ability to
transmit data from an individual’s collar to satellite has improved the scope and efficiency
of field-based research, allowing collection of the best possible data on home ranges,
seasonal movements, human-wildlife interaction zones, migration routes and speeds of
migration.
Across Africa, GPS satellite tracking has become increasingly important and accessible
during the past decade. Collaring is supported by most African government conservation
agencies and NGOs such as the Wildlife Conservation Society, Save the Elephants, WWF,
Conservation International, IUCN, National Geographic and the Born Free Foundation.
The first trial of satellite collars in Africa was on desert-dwelling elephant in the Kunene
Region, Namibia, in 1986 (Lindeque, 1991; Lindeque & Lindeque 1991). The reduced
intensity of monitoring and large number of high-resolution locations obtained by satellite
tracking provided a greater understanding of elephant movements and ranges than ever
before. African countries where GPS satellite collars are currently deployed for wildlife
tracking include Kenya, South Africa, Botswana, Mozambique, Namibia, Mali and
Gambia.
Tracking of giraffe has historically been limited to direct field observation (e.g. Foster &
Dagg, 1972; Berry, 1978; van der Jeugd & Prins, 2000), although several studies using
telemetry have provided considerably more data and greater accuracy than incidental
observations and monitoring (e.g. Langman, 1973; Dagg & Foster, 1982; Scheepers,
1992). Tracking studies of giraffe continue to require intensive field monitoring, although
finding and following individuals is considerably easier, more reliable and time efficient
than relying on chance encounters. However, no GPS satellite collaring of giraffe has
previously been undertaken in Africa.
In this chapter, direct observations and GPS satellite technology were used to provide a
better understanding of the home ranges and seasonal movements of giraffe. Importantly,
it was anticipated that the knowledge obtained could be used to protect giraffe populations
Chapter 5: Home Range
98
in the communal areas of the study region and the broader Kunene, where giraffe
conservation is currently limited.
5.1.1. Aims
The movements and ranges of individuals within a population provide insight into their
behavioural activities and ability to track resources, as well as the area needed to sustain
their ecological and biological requirements. In order to examine these issues, in this
chapter I investigated:
•
home range size of giraffe;
•
movements (hourly, daily, monthly, seasonal and long distance); and
•
movement corridors of giraffe in the three study populations (Hoanib, Hoarusib
and Khumib Rivers).
Each study area was treated as a separate unit, although movements between them were
observed.
The investigations were intended to provide an overview of the range
requirements and the dispersion of giraffe throughout the study region, and also the
movement of giraffe in relation to temperature.
5.2. Methods and materials
Over a two-year period (2002 and 2003), monthly field trips were conducted in the study
region and observations of giraffe were made predominantly by vehicle and sometimes on
foot. The Hoanib and Hoarusib River study areas were visited more frequently than the
Khumib River study area due to time, distance and logistical constraints. Once spotted,
locations and co-ordinates of individual giraffe were recorded using a Garmin II plus
Global Positioning System (GPS) which was fixed to the vehicle with an external antenna.
These co-ordinates, along with relevant bio-data (see section 4.2), were transferred to a
Microsoft Excel Professional 2000 database and MapInfo Professional 6.5 Geographical
Information System (GIS) (MapInfo Corporation, 2001). Additional data were collected
from GPS satellite collars fitted to four giraffe (see section 5.2.2). Movement corridors
Chapter 5: Home Range
99
were established by overlaying GPS co-ordinates for giraffe onto dedicated MapInfo maps
of the study area.
5.2.1. GPS satellite collaring
The satellite system used for GPS tracking during the study was an Inmarsat 3, the first in a
series of five third-generation satellites (see Appendix 8). Using spot-beam technology to
supply data communication services, a navigation transponder enhanced the accuracy,
availability and integrity of the GPS and Glonass satellite navigation systems.
The GPS unit (MT2000 terminal) was a lightweight (<900 g) mobile transceiver designed
to utilise L-band satellite frequencies.
The unit was compact with a single sealed
enclosure, constructed from chemical-resistant polycarbonate, with an integral antenna
subsystem and transceiver for communicating with satellites. A top antenna cover was
welded to a bottom shell, while gaskets and weatherproof connectors sealed the unit. The
unit was powered by an external battery pack consisting of six ‘DD’ batteries positioned on
the opposite side of the collar to the GPS unit in a protective casing similar to that used for
the GPS unit. The batteries and GPS unit were connected via insulated and protected
wiring. The battery pack also served as a counter balance for the collar, enabling the GPS
unit to remain upright.
Each of the GPS units also had an in-built VHF telemetry
transmitter that enabled tracking of individual giraffe using a hand-held antenna
(frequencies ranged from 148.150 to 148.360 MHz). The GPS collars were designed and
constructed by African Wildlife Tracking Services, South Africa.
Four giraffe (one cow and three bulls) in the study region were fitted with GPS satellite
collars during a three-day collaring period in September 2002 (28–30/09/02) (see
Appendix 9). Two collars were fitted to bulls in the Hoarusib River study area. In the
Hoanib River study area collars were fitted to one bull and one cow (Figure 5.1). The
collaring was undertaken in accordance with the MET research/collecting permit number
497/2001 and under the auspices of the Namibian Elephant & Giraffe Trust (NEGT). The
collaring team consisted of myself, a wildlife veterinarian, a wildlife health advisor, an
Chapter 5: Home Range
100
MET game capture specialist, MET staff, donors, support staff and communal conservancy
members.
Figure 5.1. Example of GPS satellite collared giraffe in the study region: collared cow
giraffe in the Hoanib River study area, September 2003 (photo courtesy of R. Caudle).
Giraffe were darted from a Bell Jet Ranger helicopter with A3080; an immobilisation drug.
The dosage used to sedate the giraffe during the collaring ranged from 13 to 16 mg of
A3080 dependent on the size and sex of the individual. Once the drug took effect, the
giraffe were brought to the ground by the game capture specialist and support staff. Once
restrained on the ground the giraffe were held down by hand and the reversal drug, M5050
(diprenorphine), injected. The giraffe were also blindfolded and ears blocked during the
collaring procedure to reduce stress. Whilst the collar was fitted, blood and tissue samples
Chapter 5: Home Range
101
(see Chapter 3) were taken, and measurements of the giraffe were collected for other
studies. The mean down time for the four collared giraffe was six minutes and no injuries
or post-capture stress were observed. All giraffe were monitored intensively for a 24-hour
period to ensure no lasting effects of the drug or collaring.
All GPS satellite collars were programmed to transmit at eight-hour intervals, providing
three GPS readings and ambient temperatures per day. The collars were expected to
transmit data for up to two years using eight-hour transmitting intervals.
However,
complications, such as collars twisting and breaking wiring connections, were experienced
and the life span of no collars exceeded 3.5 months. In addition, three of the four collars
were programmed to transmit hourly GPS readings for 4 to 5 consecutive days per month
to obtain high resolution data for hourly, daily, weekly and monthly analyses.
Mean monthly and weekly ranges were estimated using 100% MCP. Mean hourly and
daily movements were calculated in the MapInfo extension program Range Manager v.1
by tallying total distance travelled.
5.2.2. Home range analysis
Estimates of home range size were calculated using data from individually identified
giraffe observed during field research and, where available, data from the GPS satellite
collars. Home range sizes were estimated using both the 100% minimum convex polygon
(MCP) and 95% or peeled MCP methods (Jenrich & Turner, 1969). These methods were
used to enable comparisons with other giraffe studies in Africa. The MCP 95% method is
considered to provide a better estimate of home range as it reduces outlier fixes
representing forays or other potentially random movements (Broomhall et al., 2003).
The home range estimates were obtained using the MapInfo extension program Range
Manager v.1 (Data Solutions, 1998). Range Manager was developed as shareware for the
display and analysis of animal location data in the MapInfo Desktop GIS Mapping System
(MapInfo Corporation, 2001).
Chapter 5: Home Range
102
Individual home ranges were estimated for giraffe with at least 10 observations collected
over a minimum period of one year.
Mean home range sizes were calculated for
individuals within each of the three study areas as well as for bulls, cows and juveniles
throughout the study region. Published estimates for giraffe have used fewer observations
to obtain home ranges (Foster & Dagg, 1972; Berry, 1978), however, no asymptotic home
range size has been reached for any population, as the numbers of fixes obtained per
individual were considered too low. The results obtained during this study provide the first
minimum home range size estimates for giraffe in Namibia.
Differences in home range estimates between sexes and study areas, as well as between
sexes within the study areas, were tested using Kruskal-Wallis and two-tailed MannWhitney U tests. These non-parametric techniques were preferred to analysis of variance
due to non-normality in the distribution of data, which could not be correlated by
transformation. The MCP 100% and 95% home range size estimates were compared to
ascertain differences between individuals and sexes when outlier fixes were removed.
5.3. Results
5.3.1. Home range
From direct observations, the largest recorded home range estimates were obtained for a
giraffe bull (1 950 km2) and the second largest for a cow (1 098 km2) (Table 5.1). Home
range estimates using the MCP 100% and 95% methods varied markedly both between and
within the sexes (Table 5.1; also see Appendix 10). Data on home range sizes of juvenile
giraffe were limited (n =2), but showed that juveniles occupied smaller areas than bulls or
cows.
The MCP 95% home ranges of giraffe were markedly reduced when compared to the MCP
100% estimates. The reduction in home range size between the MCP 100% and 95%
estimates was about 50% for cows and 70% for bulls and juveniles. Eliminating outlier
fixes using the MCP 95% method appeared to provide a more accurate estimate of home
range sizes and core resident areas, as it excluded areas which giraffe used infrequently,
Chapter 5: Home Range
103
such as gravel plains.
The mean home range sizes of giraffe bulls in the total study area (MCP 100% and 95%)
were greater than those of cows (two-tailed Mann-Whitney U test, U = 174; P=0.003 and
U = 208; P=0.016, respectively). In the Hoanib River study area, where sample sizes were
sufficient for comparisons, home ranges of bulls (MCP 100% and 95%) were also greater
than those of cows (two-tailed Mann-Whitney U test, U = 56; P=0.006 and U = 74;
P=0.039, respectively).
Table 5.1. Mean annual home range size estimates (km2) and ranges (km2) of giraffe in the study
region using the 100% and 95% MCP method (n = number of individuals).
Estimates of home range size (km2)
MCP 100%
Study Region
2
n
MCP 95%
Mean (km )
2
2
Range (km )
Mean (km )
Range (km2)
Bull
Hoanib River
20
494.1
26.7-1 950
330.4
11.5-1 773
Hoarusib River
22
572.6
33.9-1 627
408.5
10.1-14.9
Khumib River
2
67.1
66.1-68.1
22.7
21.8-23.7
Study Region
44
513.9
26.7-1 950
355.5
11.5-1 773
Hoanib River
13
219.7
12.9-1 098
119.1
8.3-702.1
Khumib River
3
117.3
34.6-158.6
23.6
23.5-23.9
Study Region
16
199.5
12.9-1098
100.0
8.33-702.1
2
20.8
10.7-30.9
14.5
10.3-18.7
Cow
Juvenile
Study Region
The largest mean home range estimate was obtained for giraffe bulls in the Hoarusib River
study area (572.6 km2). This was marginally larger than that in the Hoanib River study
area (494.1 km2) and substantially larger than that in the Khumib River study area (67.1
km2). However, these estimates (MCP 100% and 95%) were not significantly different
between the three areas (Kruskal-Wallis: MCP 100%; H =1.829; d.f.=2; P=0.401; and
MCP 95%; H =1.437; d.f.=2; P=0.487, respectively) because of the large variation in
individual range estimates within each area. Limited numbers of individuals and poor
access to the Khumib River study area may have contributed to an underestimate of home
range size, making comparisons between areas unreliable.
Chapter 5: Home Range
104
The mean home range estimate for cows in the Hoanib River study area (219.7 km2) was
almost twice that for cows in the Khumib River study area (117.3 km2).
However,
comparisons using MCP 100% and 95% were not significantly different (Kruskal-Wallis:
Hoanib-Khumib; H = 0.222; d.f.=1; P=0.637 and H = 0.041; d.f.=1; P=0.84, respectively)
due to the large variance in range estimates within the two study areas.
Limited
observations of individual cows in the Hoarusib River study area made any analysis of
home range size impossible, but my observations suggested that home range size there is
similar to that of cows in the Hoanib River.
5.3.2. Movements
No seasonal movements or seasonal home range partitioning was observed between the
study areas. Giraffe predominantly used the main riverbeds in all the three areas, but they
also foraged up the tributaries and sometimes moved into the adjacent study area (see
Appendix 11). Giraffe movements between the Hoanib and Hoarusib River study areas
were observed on >10 occasions, although more were apparent on analyses.
In the Khumib River study area both sexes moved year round along the main riverbed, as
little vegetation was available elsewhere. In the Hoarusib River study area bulls often
moved between the main riverbed and the southern tributaries, while cows resided for most
of the year in the southern tributaries. During the hot-dry season cows were observed in
the Gomatum River, a tributary of the Hoarusib River, for short periods before returning to
the southern tributaries of the Hoarusib. In the Hoanib River study area, most adults
moved along the main riverbed year-round, while increased movements by subadult bulls
and cows into the northern and southern tributaries were observed in the wet and cold-dry
seasons.
5.3.3. Long distance movements
Long distance movements (>50 km), often lasting between one and seven days, were most
commonly undertaken by giraffe bulls, although a small number of cows also moved long
Chapter 5: Home Range
105
distances. No correlation between long range movements and seasons was apparent for
either sex. The majority of the long distance movements were observed between the
Hoanib and Hoarusib River study areas, a distance of approximately 70 km in a straight
line.
Thirty individual giraffe bulls, representing approximately 42% of the bull population in
the study region, moved between the Hoanib or Hoarusib River study areas, and viceversa. However, only a third of these (n =10) were recorded in the actual Hoanib and
Hoarusib riverbeds. One bull moved between the Hoanib and Hoarusib River study areas
three times during the study period. Only one bull was observed in both the Hoarusib and
Khumib River study areas, and would have moved approximately 35 km in a straight line
to achieve this movement.
Seven cows were recorded undertaking long distance movements. One moved from the
Hoarusib River to the Hoanib River study area over a period of two days before returning
to the southern tributaries of the Hoarusib River. Another cow moved from the Hoanib
River study area into the southern tributaries of the Hoarusib River, a movement of some
55 km. Five cows that resided predominantly in the southern tributaries of the Hoarusib
River study area, moved occasionally into the Gomatum River before returning to the
southern tributaries of the Hoarusib River catchment, but movements were not greater than
50 km. This indicated that cows moved substantially within the Hoarusib River study area,
but avoided the Hoarusib River itself. Unfortunately, limited sightings of these five cows
over the study period precluded analysis of their home ranges. During their long distance
movements, two juveniles accompanied these five giraffe cows. In general, juvenile home
ranges appeared substantially smaller than those of bulls or cows, but limited observations
restrict the reliability of range estimates.
5.3.4. Movement corridors
A number of movement corridors were identified within and between the study areas
(Figure 5.2). Movements occurred predominantly along the main riverbeds within each of
the three study areas, while movements between the areas and into alternative forage areas
Chapter 5: Home Range
106
were undertaken via the main north-south tributaries. These movements are the first
recorded movements by giraffe between the three river systems.
Chapter 5: Home Range
107
A
B
C
D
&
Figure 5.2. GPS locations of all giraffe sightings during the study. All giraffe were assigned to the study area of their first observation: (A) Hoanib
0
12.5
25
kilometres
River; (B) Hoarusib River; and (C) Khumib River. All giraffe sightings (D)—known and assumed movement corridors.
Chapter 5: Home Range
109
5.3.5. GPS satellite collar movements
Technical problems and up-link failures limited the success of the GPS satellite collars.
None of the collars transmitted for longer than three and a half months (Table 5.3).
However, for transmitted data, the numbers of non-readings and unusable fixes were low.
A total of 732 readings were obtained for the three collared giraffe bulls and 46 for the one
cow.
Table 5.3. Number of transmitted readings and non-readings received for four GPS satellite
collared giraffe and last date of transmission (collaring dates: 28-30/09/02).
Sex
Adult bull
Adult cow
Individual
Readings
Last date of
transmission
Non-readings
HSBM3
430
16
31/12/02
HSBM12
301
11
12/01/03
HNBM17
1
1
05/10/02
HNBF18
46
0
18/10/02
Despite the limitations, the data obtained from the satellite collars provided detailed insight
into the daily, weekly, monthly and annual movements of giraffe (Table 5.4). Home range
estimates were larger for the satellite collared bulls (HSBM3 & HSBM12) than the cow
(HNBF18). Weekly, monthly and annual mean home ranges were larger for one bull
(HSBM3) than the other (HSBM12).
Table 5.4. Weekly, monthly and annual home range estimates (km2) using the MCP 100% method
for two GPS satellite collared giraffe bulls (HSBM3 & HSBM12) and one cow (HNBF18), October
2002-January 2003. N.B. Annual home range estimates include ground observations.
HSBM3a
Weekly
Mean (km2)
Range (km2)
Monthly
Mean (km2)
Range (km2)
Annual
Total (100% MCP)
Total (95% MCP)
Home range estimates (km2)
HSBM12b
HNBF18c
91.57
14.52-229.7
67.56
13.09-155.4
6.11
3.65-8.79
414.2
352.4-480.9
300.53
116.2-505.6
17.67
*
1 260
1 092
1 099
778.4
41.98
37.78
*insufficient data available; a – weekly (n = 12 weeks) and monthly (n = 3 months); b - weekly (n = 14
weeks) and monthly (n = 4 months); c - weekly (n = 3 weeks) and monthly (n = 1 month).
Chapter 5: Home Range
110
Additional GPS locations for the annual home ranges of the collared giraffe where
collected using conventional radio tracking, equating to approximately 10%.
The weekly, monthly and annual (95% MCP only) home range estimates of HSBM12 were
between 71.3% and 73.8% of the size of those observed for HSBM3. Furthermore, the
averaged monthly range estimates for both bulls were 4.5 times their mean weekly range
estimates, while the annual range estimates were 2.6 times their monthly range estimates.
Due to the limited data and sample sizes for weekly and monthly ranges, no general
conclusions regarding giraffe ranges and movements can be drawn for the study region.
However, the data obtained from the collared animals, in combination with the mean
population home range size estimates, imply that giraffe bulls had substantially larger
home ranges, as well as greater weekly and monthly ranges, than cows. Range areas
increased in both sexes over time.
5.3.6. Daily movements
Data on daily movements were obtained from hourly fixes from two satellite collared bulls
and one collared cow during the hot-dry season over a combined total of 180 days (Table
5.5).
The average daily movements of bulls were 5.64 km compared to 1.87 km for the cow.
The difference in the daily movements of the two bulls was not in proportion to the
difference in their estimated weekly, monthly and annual home ranges. This variance is a
reflection of the non-linear foraging pattern of giraffe. Giraffe commonly re-use or return
to an area to forage within a 24-hour period and/or over consecutive days.
Chapter 5: Home Range
111
Table 5.5. Daily linear movements (mean ± s.d., km) and ranges (km2) of two GPS satellite
collared giraffe bulls (HSBM3 & HSBM12) and one cow (HNBF18), October 2002-January 2003.
Month
Mean daily
(Hot-dry season)
movement (km ± s.d.)
Range (km2)
HSBM12 (n = 83)
October
7.51 ± 7.02
0.90-20.86
November
6.12 ± 6.25
0.29-28.81
December
4.31 ± 2.15
0.09-8.95
January
5.29 ± 2.83
2.08-9.58
October
4.99 ± 4.36
1.16-21.39
November
6.12 ± 6.52
0.40-32.24
December
5.16 ± 2.97
0.42-10.08
1.87 ± 2.05
0.06-7.41
HSBM3 (n = 82)
HNBF18 (n = 15)
October
N.B. although expressed as means ±s.d., the movement data are derived from repeated
measurements on the same individuals and hence cannot be considered independently. In this
context, means ±s.d. are presented simply as convenient summary statistics.
5.3.7. Hourly movements
While the GPS satellite collars provided co-ordinates of giraffe locations, ambient
temperatures were also transmitted (Figure 5.3). When analysing the data on the two bulls,
no correlation was found between daily distance travelled and the daily ambient
temperature (r = 0.129, P>0.05), nor between hourly movements and hourly ambient
temperatures (r = -0.345, P>0.05). However, hourly linear movements were shorter predawn but increased at dawn and post-dawn in correspondence with an increase in ambient
temperature. During the hottest periods of the day, giraffe travelled only short distances.
Mean hourly movements increased again pre- and post-dusk, corresponding to reduced
ambient temperatures.
Chapter 5: Home Range
112
HSBM12 – November 2002
40.0
1.00
0.80
30.0
0.60
20.0
0.40
10.0
0.20
0.0
0.00
0
4
8
12
16
20
1.00
50.0
0.80
40.0
0.60
30.0
0.40
20.0
0.20
10.0
0.00
0.0
0
4
8
12
16
Temperature °C
Distance (km)
HSBM3 – November 2002
20
HSBM3 – December 2002
1.60
50.00
1.40
40.00
1.20
1.00
30.00
0.80
20.00
0.60
0.40
10.00
0.20
0.00
0.00
0
4
8
12
16
20
Time (hour)
Figure 5.3. Hourly linear movements (km) (black blocks) and hourly temperature (°C) (dashed red
line) of two satellite collared giraffe bulls in the hot-dry season 2002. Note different distance scales
on graphs.
Chapter 5: Home Range
113
5.4. Discussion
5.4.1. Home range
Giraffe bulls in the study region had substantially larger home ranges than cows (>2.5
times). The two largest home ranges of bulls were 1 950 km2 and 1 627 km2, respectively,
and represent the largest recorded home range estimates for bulls in Africa (see Appendix
12) (Foster & Dagg, 1972; Langman, 1973; Berry, 1978; Leuthold & Leuthold, 1978; Le
Pendu & Ciofolo, 1999; Van der Jeugd & Prins, 2000). Previously, the largest reported
home range for a bull was from the desert-dwelling giraffe population of Niger (1 559 km2;
Le Pendu & Ciofolo, 1999). The two largest bull home ranges in this study incorporated
two of the main riparian woodlands in the Hoanib and Hoarusib Rivers, although both
giraffe also foraged in the tributaries and the mountains between the two study areas. The
large home range sizes reported in the study region presumably reflect the reduced
probability of giraffe encountering potential mates in the sparsely populated northern
Namib Desert. However, for both bulls, the large home ranges could represent efforts to
achieve or maintain dominance in the study population, or may reflect the movements of
two large animals seeking forage to maintain their energetic and nutritional requirements.
The home ranges of giraffe cows were substantially smaller than those of bulls, with only
one cow occupying a range >1 000 km2. The largest home range reported for a cow was in
arid Niger (1 379 km2; Le Pendu & Ciofolo, 1999).
The giraffe in Niger migrate
seasonally across communal farmland, similar to that in the northern Namib Desert, and
between two distinct foraging areas (Le Pendu & Ciofolo, 1999).
The similar
environmental conditions experienced by giraffe in Niger and Namibia help to account for
the large ranges of giraffe in both regions.
Home range sizes of juvenile giraffe were small (20.81 km2) compared to those of both
cows and bulls in the study region, however, they did correspond with their mothers’ home
ranges for the same periods. Although ranges of juvenile were greater than those observed
elsewhere (e.g. 12.8 km2 in South Africa, Langman, 1973; see Appendix 12), they still
appeared to be underestimates of the true ranges occupied.
For example, juveniles
observed in the southern tributaries of the Hoarusib River ranged further than the mean
Chapter 5: Home Range
114
estimates reported (personal observation), but low numbers of observations restricted home
range size analysis for these individuals. These juveniles were sighted at all times with
their mothers who moved between the southern tributaries of the Hoarusib River study area
and the Gomatum River in the hot-dry season, covering an estimated area of 200 km2.
Increased numbers of sightings would provide a better indication of the home ranges of
juveniles in the northern Namib Desert.
Published estimates indicate that giraffe in arid environments have relatively large home
ranges, and that bulls have larger ranges than cows (Scheepers, 1992; Le Pendu & Ciofolo,
1999). Smaller home ranges are observed for giraffe in more densely vegetated savanna
environments, while sex differences in range size are not as pronounced (Foster & Dagg,
1972; Langman, 1973; Berry, 1978; Leuthold & Leuthold, 1978; van der Jeugd & Prins,
2000; see Appendix 12). Range sizes also vary between areas where giraffe have different
population structures and densities, as well as between environments differing in forage
availability, seasonal rainfall, predators, management regime, intensity of hunting and area
size (see Appendix 12).
A negative correlation (r = -0.45; see Appendix 12) was observed between published
estimates of home range size against population density in giraffe, indicating that giraffe in
sparse populations generally occupy larger mean home ranges. Results from the study area
support the concept that low population density and large range are correlated, as reported
for arid environments generally (Scheepers, 1992; Le Pendu & Ciofolo, 1999). Decreased
availability of forage and high variability in seasonal and spatial rainfall are important
factors limiting density and increasing the range sizes of giraffe in arid environments
(Scheepers, 1992; Le Pendu & Ciofolo, 1999).
Previous research indicates that the home ranges of giraffe also vary in shape. Along
forage rich riparian environments, home ranges are often linear and elongated (Berry,
1978; Scheepers, 1992) but can be irregular in shape (Dagg & Foster, 1972; Leuthold &
Leuthold, 1978; Pellew, 1984b; du Toit, 1990a; Le Pendu & Ciofolo, 1999; van der Jeugd
& Prins, 2000). In the study region most home ranges of giraffe were linear because the
riparian environments are the lifelines of the northern Namib Desert; there is little forage
available elsewhere. Home range sizes have not been reported for giraffe in the Hoanib
Chapter 5: Home Range
115
River because of these linear movements (Scheepers, 1992). However, during this study,
giraffe in the study region were found to be more mobile than previously assumed, thus
allowing range estimates to be made. Despite some individuals moving along the northern
and southern tributaries of each catchment and between the study areas, giraffe seldom
used all the area in their home range. This is because much of the home range is devoid of
forage and encompassed inhospitable terrain.
Home ranges of giraffe bulls and cows in the study overlapped both within and between
the three study areas. In both the Hoanib and Khumib River study areas, giraffe of both
sexes foraged predominantly along the riparian woodland of the main riverbeds and their
tributaries. No apparent spatial or seasonal segregation in habitat use was observed for
either sex in the study area. However, spatial segregation between bulls and cows did
occur in the Hoarusib River study area. No cows were sighted in the Hoarusib River
during the study period, and they foraged only infrequently in the Gomatum River,
residing predominantly along the study area’s southern tributaries. The Hoarusib River
population is strongly bull-biased. The residence of bulls in the main Hoarusib and
Gomatum riverbeds can be attributed to the greater availability of food along these rivers
than elsewhere in the study area. Giraffe bulls need more food than cows to maintain their
larger body mass, hence bioenergetic advantages would presumably accrue to bulls able to
exploit the Hoarusib River.
If rich food resources were available in the Hoarusib River study area, why should bulls,
but not cows, exploit them? It has been hypothesised that, in various ungulate species,
cows forgo foraging benefits for environments that are more suitable for raising young
(Main & Coblentz, 1990). Giraffe cows in the Hoarusib River study area foraged only
seasonally in the Gomatum River during the hot-dry period when food availability was
reduced elsewhere. The avoidance of the Hoarusib and, to a lesser extent, Gomatum
Rivers by cows may be an attempt to reduce conflict with both bulls and communal
farmers in the area, and thus protect their offspring by limiting potential interactions with
bulls, people and/or villages.
Similar hypotheses have been postulated for the local
distribution of giraffe cows and their offspring in other parts of their range (e.g. Foster,
1966; Dagg & Foster, 1982; Pratt & Anderson, 1982; Pellew, 1984a; Young & Isbell,
1991; Ginnett & Demment, 1997). This resource-shelter trade-off may be feasible when
Chapter 5: Home Range
116
population numbers are low, however, if giraffe numbers increase it may begin to deplete
food resources in the tributaries and mountains away from the main riverbeds. In the longterm, this may result in cows increasing their time spent in the main riverbed during
periods other than the hot-dry season to supplement nutrient intake and maintain their
bioenergetic needs.
Sex-dependent and geographical patterns of segregation have been reported in other giraffe
populations (e.g. Foster, 1966; Pratt & Anderson, 1982; 1985; Caister et al., 2003).
However, home ranges of individuals in these populations overlapped considerably, if only
seasonally, and no permanent socio-spatial segregation or territoriality has been reported
(e.g. Backhaus, 1961; Foster & Dagg, 1972; Berry, 1978; Leuthold, 1979; Le Pendu et al.,
2000; van der Jeugd & Prins, 2000). Overlapping home ranges are common in other
ungulates such as impala (Jarman, 1972; Matson, 2003) and kob (Fischer & Linsenmair,
1999), while sex-dependent segregation in habitat use has also been reported in kudu (du
Toit, 1995) and red deer (Clutton-Brock et al., 1982). Habitat segregation can result from
many factors, although it is driven predominantly by the quality of available food—cows
have higher energetic and nutritional requirements during pregnancy and lactation (giraffe
- Ciofolo & Le Pendu, 2002; Caister et al., 2003; other ungulates - Clutton-Brock et al.,
1982; du Toit, 1995; Fischer & Linsenmair, 1999), and will exploit the richest habitats that
still provide some security for their offspring.
5.4.2. Movements and corridors
Giraffe in the study region showed the highest preference for the riparian environments,
almost to the exclusion of other habitat types. Giraffe were sighted in all other habitat
types at some stage throughout the study period, but their use of or residence time in these
habitats was low. The strong preference of giraffe for riparian environments is similar to
that reported for desert-dwelling elephants in the northern Namib Desert (Viljoen, 1988;
1989b). Giraffe and elephant spend much of their time foraging in the riparian woodlands
in both the Hoanib and Hoarusib River study areas, although elephant are absent from the
Khumib River study area.
Chapter 5: Home Range
117
Movements in each of the study areas were predominantly along the main riverbeds and
their tributaries. Giraffe were observed traversing and, on occasion, opportunistically
foraging in open gravel plains and dune habitats; one cow was seen foraging on a steep
mountain ridge. Giraffe usually avoid rocky habitats (Bond & Loffell, 2002), but may
forage there if food is scarce elsewhere. Movements of giraffe across the barren open
plains were predominantly straight-line traverses between two forage sources, and usually
between the riparian woodlands and the food that they contain.
During the cold-dry season and occasionally the early hot-dry season, giraffe were
identified venturing into and exiting the mountains between the Hoarusib and Khumib
Rivers. No designated tracks or roads exist in this remote and rocky terrain, which limits
vehicular access. However, it was assumed that giraffe were using these areas away from
the riverbed as alternative seasonal forage sources during this time of the year to
supplement their forage requirements. Precipitation in the form of evening fog blankets
the mountains and raises the average moisture content in rarely-foraged mountain plant
species, such as Commiphora, Maerua and Boscia spp. (Viljoen, 1988; Scheepers, 1992).
Selective browsing of these plant species is considered to be the reason giraffe venture into
the mountains to forage at certain times of the year. Desert-dwelling elephant have also
been observed in the mountains at the same time, foraging on the same moisture-rich plant
species (Viljoen, 1988; 1989b; K. Leggett, personal communication; personal observation).
While distinct seasonal movements of giraffe between areas are limited, short excursions
into the seasonally forage-rich mountain areas occur. The desert-dwelling giraffe in Niger
have dry season home ranges twice the size of those during the rainy season due to the
seasonal movement of giraffe between two distinct forage areas (Ciofolo & Le Pendu,
2002; Caister et al., 2003).
In general, movements and home range sizes of giraffe have been linked strongly to
seasonal browsing and/or the availability of water (Foster & Dagg, 1972; Leuthold &
Leuthold, 1972; Hall-Martin, 1974; Berry, 1978; Leuthold & Leuthold, 1978; Pellew,
1984b; Scheepers, 1992; Le Pendu et al., 2000; van der Jeugd & Prins, 2000; Caister et al.,
2003).
Specifically, seasonal movements of giraffe have been associated with
phenological changes in preferred plant species (Hall-Martin & Basson, 1975; Leuthold &
Leuthold, 1978; Scheepers, 1992; Le Pendu & Ciofolo, 1999), with shifts in preferences
Chapter 5: Home Range
118
for plant species leading to seasonal expansion or contraction of ranges (Leuthold &
Leuthold, 1972; Hall-Martin & Basson, 1975; Scheepers, 1992; Le Pendu et al., 2002).
The few observations of giraffe drinking in the study region (less than ten sightings in 70
years: Viljoen, 1981; Scheepers, 1992; Fennessy et al., 2003) suggest that surface water is
not a factor influencing either seasonal or annual movements of giraffe, as they do not
actively seek this resource out.
Seasonal movements of giraffe east-west and associated use of the riverbed, have been
reported in the Hoanib River study area (Fennessy et al., 2003). During the hot-dry season
giraffe aggregate there in response to the increased availability of Faidherbia albida pods,
an essential dry-season food, and an early flush in Colophospermum mopane leaves.
Acacia species, Salvadora persica and Euclea pseudebenus are also relatively abundant,
but little other vegetation is available along the river. During the hot-dry season giraffe are
more frequently observed in the eastern and central parts of the Hoanib River study area.
This may be a response to F. albida podding earlier here than in the western section,
raising the question of whether giraffe move in relation to forage availability, forage
nutrients and moisture composition, or other additional factors. Chemical analysis of the
preferred forage of giraffe along an east-west gradient in the Hoanib River is discussed in
Chapter 7, with the findings posited to provide a better understanding of why subtle
seasonal movements of giraffe occur along this riparian strip.
5.4.3. Long distance movements
This study recorded the first movements of giraffe between any of the three study areas.
Giraffe were previously assumed to reside in only one catchment and to be restricted to the
main riverbed and its tributaries (Scheepers, 1992). The large distances between study
areas across the arid northern Namib Desert were assumed to inhibit any large-scale
movements by giraffe (Viljoen, 1982; Tarr, 1986; Scheepers, 1992).
The initial
observation of giraffe movement between the study areas was recorded for a GPS satellitecollared bull giraffe (HNBM17). This specific unit transmitted only one usable reading
following collar placement. The bull moved north from its collaring site in the Hoanib
River into the Hoarusib River study area, a direct line distance of 70 km. However, due to
Chapter 5: Home Range
119
the topography of the area, the movement would more likely equate to a distance of about
80 km; this was confirmed by ground truthing of the bull’s movement. This initial finding
provided insight not only into the magnitude of giraffe movements and home ranges, but
also population structure and gene flow in the study area. Throughout the remainder of the
study, increased numbers of identified giraffe were observed traversing the Hoanib and
Hoarusib River study areas.
Many giraffe bulls made long distance movements between study areas (42% of
individuals), similar to bulls studied in a low-density arid area in Niger (Le Pendu &
Ciofolo, 1999). Long range movements are probably the result of numerous factors such
as searching for receptive cows, seeking or maintaining dominance or forage, conflict with
communal farmers, tourism and local poaching. Most likely, however, long distance
movements in this study were the result of searching for receptive cows and the movement
by giraffe away from newly occupied communal settlements. Similar interactions have
been reported in Niger, with giraffe movements in both areas synchronised with communal
farmer activities (Ciofolo, 1995; Le Pendu & Ciofolo, 1999). Long range movements of
giraffe have been observed elsewhere in Africa but have been attributed usually to
searching for forage or bulls seeking receptive cows (Berry, 1978; Dagg & Foster, 1982).
Only one giraffe, a bull, was recorded moving between the Hoarusib River and Khumib
River study areas, a distance of less than 50 km. This bull was not observed associating
with any giraffe in the Khumib River before returning to the Hoarusib River study area,
although it is possible it may have been in search of receptive cows. Its movement into the
Khumib River area was not correlated with any known human-wildlife conflict issues and
the bull spent much of its time there in close proximity to the SCP WSN camp. It
remained in the Hoarusib River study area after returning from its excursion to the Khumib
River. As movements between areas occur, it is possible that excess bulls migrate from the
Khumib River into the Hoarusib River area to relieve pressure on forage.
It is considered that most long distance movements by giraffe cows in the study area
reflected searches for seasonal forage. Movements between the southern tributaries of
Hoarusib River and the Gomatum River were observed during the hot-dry season. Cows
were observed in the Gomatum River only during the hot-dry season for brief periods
Chapter 5: Home Range
120
before returning south to the tributaries between the Hoarusib and Hoanib River study
areas. Movements of cows along the Hoanib River were predominantly into its northern
tributaries (Swaragab and Obias Rivers) where forage was available. Only one cow moved
between both the Hoanib and Gomatum riverbeds and it was often sighted feeding along
the tributaries between the Hoanib and Hoarusib River study areas. Its behaviour was
similar to that of a dominant bull searching for receptive cows; it associated infrequently
with any other giraffe, independent of sex or age class, and spent little time in any one
location.
Research on elephants in the northern Namib Desert has correlated long-range movements
with forage quality, proximity of forage to water, seasonal rainfall and flood events
(Cooper, 1980; Viljoen, 1980; Tarr & Tarr, 1989; Viljoen & Bothma, 1990a). Elephants
have been observed moving up to 70 km in a 24-hour period between forage and water
resources in the northern Namib Desert (Viljoen & Bothma, 1990a; Leggett et al., 2003a)
and have been recorded travelling in excess of 600 km in one season (Leggett, 2004).
However, survival of giraffe in the northern Namib Desert is not dependent on free-water,
as observed for elephant, and therefore their ranges were not correlated with proximity of
forage to water.
5.4.4. GPS satellite collar movements
The monthly range of the collared bulls was substantially larger than that of the collared
cow (more than 17 times) and greater than the average derived from direct observations.
This may suggest that the extent of movement by giraffe has been systematically
underestimated in observational studies, although there are no other comparable data for
collared giraffe. The monthly range estimates were approximately 38% of the annual
range estimates (95% MCP) for both bulls, suggesting that ranges drift across the year.
This hypothesis is further supported by the weekly range estimates being approximately
22% of the monthly range estimates for both bulls. Movements of bulls were probably
directed at seeking out nutrient-rich forage during the hot-dry season and conflict
avoidance with communal seasonal settlements. The range estimate for the collared cow in
Chapter 5: Home Range
121
the hot-dry season was small and indicated that foraging was restricted to the Hoanib River
riparian environment and its immediate surrounds.
Daily movements varied markedly between all collared giraffe. The giraffe cow’s average
daily movement and range (1.87 km) was lower than for cows elsewhere: 2.9 km in South
Africa (Langman, 1973) and ≤2.3 km in Zambia (Berry, 1978). However, the cow’s home
range (199.51 km2) was substantially larger than that of giraffe in South Africa (24.6 km2:
Langman, 1973) and Zambia (68 km2: Berry, 1978). These giraffe occupy more densely
vegetated habitats and use a greater portion of their seasonal range each day than giraffe in
open, arid environments. Similar patterns have been reported for other species in the
Artiodactyla family (Clutton-Brock et al., 1989; Fischer & Linsenmair, 1999).
The daily movement of the two collared bulls in the study region averaged 5.64 km,
similar to or greater than that observed elsewhere in Africa: 2.6 km and 5.9 km in South
Africa (Innis, 1958; and Langman, 1973, respectively) and 3 km in Zambia (Berry, 1978).
The daily movement data were collected during the hot-dry season, coinciding with the
highest ambient temperatures, which may have resulted in smaller movements. Giraffe
were observed to browse in a non-linear fashion, although they often re-visited the same
area to forage, and made intensive use of the riparian woodlands.
As a consequence of the apparently selective foraging strategy of giraffe, it has been
suggested that some plants produce increased tannin levels when eaten, preventing
browsers from feeding at one source for any length of time (Leuthold & Leuthold, 1972;
Hall-Martin, 1974; 1975; Hall-Martin & Basson, 1975; Sauer et al., 1982; Pellew, 1984a;
Caister et al., 2003). As plant tannin concentrations usually differ between habitats (Sauer
et al., 1982), it is possible that plant tannin levels in the northern Namib Desert are lower
than elsewhere, enabling giraffe to forage longer on individual trees and thus reduce their
daily movements. However, while daily movements may be short, the weekly, monthly
and annual movements of the GPS-collared giraffe were considerably larger than those
observed elsewhere. The arid conditions in the study region presumably require bulls to
shift as forage conditions change seasonally, as well as to spend more time searching for
potential partners in the sparse desert population (du Toit, 1990a).
Chapter 5: Home Range
122
The mean hourly movements of the collared giraffe in the hot-dry season were strongly
biphasic throughout 24 hours. Increased movements occurred post-dawn/early morning
and in the period following, as well as pre-dusk/early evening. Distance travelled was least
during the hottest period of the day (midday), thus reducing excess heat loads, as
postulated in the ‘heat-load’ concept (Leuthold & Leuthold, 1978; Pellew, 1984a).
Although no direct correlation was recorded between ambient temperature and distance
travelled, the results suggest that giraffe rest predominantly during the middle of the day
and show another reduced period of activity again in the late evening/early morning.
Evaluation of giraffe activity budgets (Chapter 6) provides further insight into the biphasic
movement patterns of giraffe, and their engagement in energy consuming activities at
different ambient temperatures.
The GPS satellite collars collected substantial amounts of continuous data when
functioning and provided high resolution data in a shorter period of time than is possible
using either radio collars or individual identification methods (Langman, 1973; Dagg &
Foster, 1982; Scheepers, 1992). The collars also allowed this study to record the first
detailed hourly movements of giraffe over 24-hour periods.
5.5. Methodological problems
The GPS satellite collars impacted minimally on giraffe behaviour with no signs of postcapture stress. However, the longevity of the GPS units was poor, with a maximum of
three and half months of data collected from what was anticipated to be a two year study.
The exact reasons for the technical failure are unknown as the collars will be recovered in
late 2004/early 2005. However, failure in the wiring or its connection between the battery
pack and GPS unit is assumed to be the most likely reason.
In addition to the GPS units failing to transmit for the intended period, it was noted that
collars shifted or moved around the neck of each giraffe. This was not anticipated. The
continual shifting of the collars resulted in the GPS units not remaining in an upright
position on the giraffe’s backs. The battery packs, which were intended to act as counterweights to stabilise and keep the GPS units in an upright position, often wedged
Chapter 5: Home Range
123
themselves behind the giraffe’s front legs when shifting. The battery packs never impeded
movement, but the collar shifting may have reduced the transmission quality. It was not
uncommon to observe a collar in three different positions (GPS upright, right and left) over
a 12-hour period. The more physical behaviour of giraffe bulls may have contributed to
their collars shifting more often. Initially, the shifting of the collars was assumed to be the
main reason for the lack of data transmission (D. Okuysen, personal communication),
however, the two collars that remained functional the longest were those on the giraffe
bulls, and both were observed shifting continuously.
When it worked, radio tracking assisted greatly in finding individuals in the rugged terrain,
especially those giraffe that had moved out of the riverbeds. Furthermore, the collared
individuals became ‘Judas’ giraffe in the sense that, when one was found, numerous other
giraffe were often detected due to their gregarious nature.
It is strongly recommended that design flaws are rectified and trialing of GPS satellite
collars be undertaken prior to further field use of this technology on giraffe. Testing
variations of the collar and transmitter on captive or habituated individuals would improve
cost effectiveness, field efficiency, reliability and longevity. Other features that should be
assessed in the future include the use of a data logger that could store readings for an
extended period of time before downloading to a beacon or vehicle station, hence
improving data security. Furthermore, increased battery numbers or improved battery
technology would allow for increased transmission time and data collection.
5.6. Conclusion
Estimates of home range size in the study population were on average larger than those in
other studied populations, with the exception of the desert-dwelling giraffe in Niger. In
particular, the largest individual home range of a giraffe bull was recorded during this
study. The larger home range size in my study region was correlated with low population
density, reduced forage density and increased searching for receptive cows by bulls. The
predominant pattern of movement was linear, along the riparian environments in each of
the three study areas; however, large-scale irregular movements into tributaries and other
Chapter 5: Home Range
124
areas were also recorded. Seasonal movements of giraffe in the study population were not
as distinctive as those in other giraffe populations. Small-scale movements by bulls into
the mountains above the Hoarusib River, as well as by cows into the northern tributaries of
the Hoanib River, were observed. Small-scale habitat segregation was observed in the
Hoarusib River study area with giraffe cows foraging only in the Gomatum River during
the hot-dry season; this is considered to have limited conflict with communal farmers and
tourists.
Numerous observations of giraffe, predominantly bulls, moving between the study areas
were recorded, indicating that giraffe were not as restricted in range as previously assumed.
This supports the idea that gene flow should occur between the riparian populations,
supporting the results on genetic architecture and haplotype distribution presented in
Chapter 3. Finally, use of GPS satellite collars provided some of the highest resolution
data on giraffe movements to date. Strong biphasic movement behaviour of giraffe over
24-hour periods indicated that activity correlated with ambient temperatures.
This
behaviour is further investigated in the next chapter when assessing hourly and seasonal
activity.
Chapter 5: Home Range
125